Question: In an attempt to clean up the wiring in my mini-tower desktop case and make it less of a mess to handle, I’m considering swapping out the longer SATA cables in it for ones that are only as short as they need to be. The market is full of very short SATA cables:
However, the following Seagate article states:
?
Serial ATA cables are available in many lengths up to 1 meter. Minimum cable length is 12 inches, using shorter cables can cause timing, or noise interference on the cable.
The above is the only source I’ve been able to find for the claim that SATA cables less than 12 inches will result in performance issues. Neither of the famous Maximum PC or Puget Systems SATA cable benchmarks take cables less than 12 inches into account.
For those who have had experience with these shorter cables – have you noticed any particular performance problems with them?
Even better, can anyone link to (or conduct) benchmarks that could shed some empirical light onto this claim?
Answer: The Serial ATA Revision 3.0 Specification makes no mention of minimum cable length that I could find, although the shortest cable length referenced in the standard is 1 meter.
The only reference I can find supporting the claim that shorter cables can cause timing or noise interference is from a paper entitled Successful SAS/SATA Equipment Design and Development: What You Need to Know for 6 Gb/s and Beyond:
If a cable is too short, however, de-emphasis and equalization may actually end up overcompensating for a signal, actually reducing signal integrity.
However, further reading of the paper makes it clear that this only applies to SAS-2 cables, not SATA (note emphasis):
Don t use too short a cable for SAS-2:
While this may ?seem to contradict the previous point, SAS-2 employsde-emphasis and equalization to attempt to overcome the effects ofattenuation and jitter. However, as endpoints don t communicate as tohow much de-emphasis and equalization there should be, these valuesare fixed within a range ?with expectation that they will work overthe specification s supported cable lengths.
If a cable is too short, however, de-emphasis and equalization may actually end up ?overcompensating for a signal, actually reducingsignal integrity.
While the specification supposedly can work ?for any range of cablelength, if you encounter issues with ?short cables working with SAS-2,try experimenting with ?longer lengths to determine if this is theproblem. Note that short cable issues do not apply to SATA as SATAdoes not support de-emphasis or equalization.
Further, it seems that for SATA the shorter the cable the better:
Other ways developers can improve signal integrity include:
…
Use the shortest reasonable cable:
The longer the cable, ?the greater the attenuation. The best way toreduce attenuation is by shortening cables. This is especially truefor ?SATA, where the absence of the SAS specification s de-emphasisand equalization mean that the link s characteristics completelycontrol signal attenuation and jitter.
In conclusion: despite cables shorter than 1 metre not being part of the SATA specification, the shortness of a SATA cable does not inhibit its performance, and may even improve it.
Further Reading
Successful SAS/SATA Equipment Design and Development
Serial ATA Revision 3.0 Specification